Draft version December 10, 2018 Typeset using LATEX twocolumn style in AASTeX62

Rotating halo traced by the K giant stars from LAMOST and Gaia

Hao Tian1 (LAMOST Fellow) Chao Liu,1 Yan Xu,1 and Xiangxiang Xue1

1Key Laboratory of Optical Astronomy, National Astronomical Observatories, Chinese Academy of Sciences, Datun Road 20A, Beijing 100012, PR China;

(Received December 10, 2018; Revised December 10, 2018; Accepted —–) Submitted to —-

ABSTRACT With the help of Gaia DR2, we are able to obtain the full 6-D phase space information for stars from LAMOST DR5. With high precision of position, velocity, and metallicity, the rotation of the local stellar halo is presented using the K giant stars with [Fe/H]< −1 dex within 4 kpc from the Sun. By fitting the rotational velocity distribution with three Gaussian components, stellar halo, disk, and a +4 −1 counter-rotating hot population, we find that the local halo progradely rotates with VT = +27−5 km s −1 providing the local standard of rest velocity of VLSR = 232 km s . Meanwhile, we obtain the dispersion +4 −1 of rotational velocity is σT = 72−4 km s . Although the rotational velocity strongly depends on the choice of VLSR, the trend of prograde rotation is substantial even when VLSR is set at as low as 220 km s−1. Moreover, we derive the rotation for subsamples with different metallicities and find that the rotational velocity is essentially not correlated with [Fe/H]. This may hint a secular evolution origin of the prograde rotation. It shows that the metallicity of the progradely rotating halo is peaked within -1.9<[Fe/H]<-1.6 without considering the selection effect. We also find a small fraction of counter- rotating stars with larger dispersion and lower metallicity. Finally, the disk component rotates with +6 −1 +3 −1 VT = +182−6 km s and σT = 45−3 km s , which is quite consistent with the metal-weak thick disk population.

Keywords: galaxies: individual (Milky Way) — Galaxy: kinematics and dynamics — Galaxy: halo — Stars: individual (K-giants)

1. INTRODUCTION stream, discovered by Ibata et al.(1994), many streams It has been widely accepted that the Milky Way, at have been discovered with matched filter method (Rock- least the halo, was formed through hierarchically merg- osi et al. 2002; Grillmair & Carlin 2016; Grillmair & ing smaller stellar systems, such as globular clusters and Dionatos 2006a,b,c; Grillmair 2009) and other methods (Mateu et al. 2018). But the streams are not the final

arXiv:1805.08326v2 [astro-ph.GA] 7 Dec 2018 dwarf galaxies (Belokurov et al. 2006). With a long time evolving, the stellar members of those merged sys- fate of those merged clusters or dwarf galaxies, a longer tems will be continuously stripped and then form stellar evolving time will make the streams more and more re- streams (Johnston et al. 1999). Such phenomena have laxed and finally become part of the smoothing halo, been widely searched with photometric sky surveys, e.g. i.e. it is not visible in the geometric space (Helmi et SDSS (York et al. 2000) and Pan-STARRS (Bernard al. 1999), even blurred in the phase space (Helmi & de et al. 2016). Since the first stellar stream, Sagittarius Zeeuw 2000; Sanderson et al. 2015). Assuming those merged stellar systems fell in with random orbits (Sanderson et al. 2015), the contribu- Corresponding author: Chao Liu tion of z−axis angular momentum Lz overall should [email protected] be around 0. The net angular mometum is directly related to the merging history, not only including the 2 Tian et al. minor mergers of the globular cluster and dwarf galax- mates led to a debate (Sch¨onrich et al. 2011; Beers et al. ies, such as the Sagittarius Stream, but also includ- 2012). This will be highly improved by the Gaia mission ing the major merger events, such as the recently dis- (Perryman et al. 2001). covered Gaia-Enceladus(Helmi et al. 2018), also named GDR2 (Gaia Collaboration et al. 2018a,b,c,d,e,f) con- Gaia-Sausage(Belokurov et al. 2018), which contibutes tains ∼1.3 billion stars high accuracy proper motions a larger fraction of retrogradely rotating stars. Another and parallaxes, which have brought a series of amazing possibility is that the merged halo was driven to slowly results about the halo, such as the velocity determina- rotate by the non-axisymmetric structure in the disk tion of globular clusters and rotation curve of the Large (e.g. bar) in a long time scale (Athanassoula et al. 2013). Magellanic Cloud (Gaia Collaboration et al. 2018e). Be- The studies on dynamics of the halo have been done sides, Gaia Collaboration et al.(2018e) also derived a since 1980s (Frenk & White 1980) who used 66 clus- lower limit total mass for the Milky Way. ters to constrained the rotation velocity 60±26 km s−1. Guoshoujing Telescope (the Large Sky Area Multi- As the velocity ellipsoid is directly related to the mass Object Fiber Spectroscopic Telescope, hereafter LAM- distribution of the halo (Binney & Tremaine 2008), OST) is a 4-meter, quasi-meridian, reflecting Schmidt more work focused on the distribution of the dispersions telescope which makes it an efficient spectroscopic tele- (Sommer-Larsen et al. 1997; Sirko et al. 2004; Smith et scope (Cui et al. 2012; Deng et al. 2012; Zhao et al. al. 2009; Hattori et al. 2013; Kafle et al. 2013; Loebman 2012). Liu et al.(2017) and Xu et al.(2018) used metal et al. 2018). poor K giants from LAMOST to trace the density pro- Because of the difficulty of observation of proper mo- file of the halo and found an inner oblate, outer nearly tions, especially for distant stars. Morrison et al.(1990) spherical profile. Recently, LAMOST has released its studied the kinematics of the halo and disk using the 5th data (DR5). As low resolution of R = 1800, the G and K giant stars within 4 kpc from the Sun, and uncertainties of Teff and [Fe/H] are typical ∼ 100 K and −1 claimed a prograde rotation of VT = 25 ± 15 km s . ∼ 0.1 dex, respectively, for the stellar spectra with signal They also showed that the metallicity of the thick disk to noise ratio larger than 30. stars can be as low as -1.6. Smith et al.(2009) studied The paper studies the rotation of the local stellar halo the kinematics of the local halo using ∼1700 solar neigh- using the conbination of the GDR2 and LAMOST DR5 bourhood subdwarfs and found that the stellar halo ex- K-giant stars and is constructed as follow. In Section2, hibits no rotation and the rotational velocity Vφ disper- the sample stars are selected. The model of the distibu- sion 82 ± 2 km s−1 , what should be noticed is that the tion of rotational velociy is developed in Section3. The typical velocity errors are large, about 30 to 50 km s−1. results are discussed in Section4. The conclusions are More recently, the progress of the observations and listed in Section5. data analysis techniques, high precision proper motions have been obtained through combining catalogues ob- 2. DATA SAMPLE tained at different epochs cross a long time baseline In this work, we use K giant stars selected from LAM- (Tian et al. 2017; Belokurov et al. 2018). Deason et al. OST DR5 based on the criteria suggested by Liu et al. (2017) used the 3D velocity calculated from the proper (2014) to study the rotation of the nearby halo. After motion of different tracers, including RR Lyrae, Blue cross-matching with GDR2, we obtained the parallax, Horizontal Branch (BHB) and K giant stars, in a distant proper motions, radial velocities and metallicities for the volume from Gaia Data Release (hereafter GDR) 1 and tracer stars. A robust metallicity cut with [Fe/H]< −1 SDSS dataset to study the halo rotation. A prograde is used to remove the majority of disk stars, which leaves rotation was found with V = 14±2±10 km s−1. What T 6660 K giant stars. In this case most of the thin disk is more, the metal richer stars are showing a stronger stars and many thick disk stars will be removed. As we prograde rotation which is explained as a small fraction will discuss in the following sections, the left disk com- of halo stars formed in situ. Kafle et al.(2017) also ponent will be considered during the analysis. With the found a similar trend with samples of Main Sequence distances from Bailer-Jones et al.(2018), the samples Turnoff and K giant stars only with radial velocity aval- within 4 kpc and relative distance error D /D < 0.2 ible. What is different is that the metal poorer K giant error are further selected so that the distance estimates are stars show a retrograde rotation. reliable, and 3827 stars are left in the sample. Com- Since Carollo et al.(2007) discovered the inner and paring the radial velocities between GDR2 and LAM- outer halo and claimed that the inner halo is rotating OST, we find an offset RV − RV of 5.38 km progradely, in contrast the outer components is rotating Gaia LAMOST s−1, which was also mentioned by Tian et al.(2015) and retrogradely. But the systematic bias in distance esti- Sch¨onrich & Aumer(2017). To avoid this systematics, Halo Rotation 3

0.8 103 pmra error 400 pmdec error 0.6 )

300 * ( t

M 2 n D 0.4 10 200 u o c count(*) 0.2 100

0 1 0.0 10 0 1 2 3 4 0.00 0.04 0.08 0.12 0.16 0.20 0 2 4 DM proper motion error RV

Figure 1. Distributions of the errors of proper motions, radial velocities and cylindrical velocities in the left, middle and right panels respectively. The red line in the left panel indicates the limit of the distance error 20%. The arrows indicate the median errors for each quantities.

by the cut in [Fe/H]. The top panels show that the stars are rarely sampled at X > 0, Y > 0, and Z ∼ 0, which is caused by the observational limit of LAMOST survey. Z Z − − This would not affect the detection of the rotation of the

− − halo. The bottom-left panel shows that the distribution − − − X Y in VR versus VZ space is quite smooth. And the bottom-

right panel displays an excess with VT at around 200 km −1 s , which may be the slight contamination of the thick

Z Z

V V disk. − − In Figure3, noobvious correlation between [ F e/H] − − and R is found. The colors of the dots indicate the − − GC − − distance errors, which also show no significant correla- VR VT tion with [F e/H]. This implies that the samples do not Figure 2. Distributions of the sample in geometric space experience any selection bias in metallicity. (top panels) and velocity space (bottom panels). In the middle panel, the distribution of the rotation velocity VT versus metallicity [F e/H] is shown with the we adopt the radial velocity provided by GDR2. Fig- gray dots. Clearly we find that more stars are progradely ure1 shows the error distributions of distance, proper rotating in the metal richer side, which is suggested to motions and radial velocity in the left, middle and right be the disk component. The black lines in the blue and panels respectively. We can find that the errors of dis- red shades are the rotation velocity for the halo and disk tance increase with distances. From the middle panel, components respectively, which will be discussed in the the median errors of proper motion in RA and DEC later sections. −1 are 0.065 and 0.054 mas yr , respectively, which are In the right panel, the errors of the velocities are labelled with red and green arrows. The median error shown, which are calculated following the calculation of of radial velocity is 1.0 km s−1 labelled with arrow in Johnson & Soderblom(1987). The median errors of VR, the right panel. −1 VT and VZ are 5.4, 11.3 and 4.3 km s as represented We use galpy (Bovy 2015) to calculate the 3D loca- by the red, green and blue arrows, respectively. tions in heliocentric cartesian coordinates, X, Y , Z, and velocity components in galactocentric cylindrical coordi- 3. MODEL OF THE DISTRIBUTION OF VT nates, radial component V , rotational (azimuthal) com- R In Figure4, it is seen that the distribution of V for ponent V , and vertical component V . By default, we T T Z the samples (black dots with error bars) shows two peaks adopt the velocity of the local standard of rest (LSR) and a long tail beyond V < −200 km s−1. This hints with respect to the Galactic center as V = 232 km T LSR that it may contain three components, The most dom- s−1 (McMillan 2017) and the solar motion with respect inated component shows up a peak located just at the to the LSR as (U ,V ,W ) =(11.1, 12.24, 7.25) km s−1 right side of V = 0. The second component contributes (Sch¨onrich et al. 2010). T a metal-rich peak at around 200 km s−1. And the third Figure2 shows the distributions of the selected K gi- one, maybe a broad one, contributes to the long tail at ant stars in geometric (top panels) and velocity (bottom V < −200 km s−1, especially at metal poor end. By panels) spaces. The samples do not show disk-like fea- T convenience, we assume all the three components are tures, implying that most of the disk stars are removed Gaussians and their distribution of VT can be defined as 4 Tian et al.

1.0 0.8 103

VR

200 V 0.6 T VZ 1.5 ) * ] ( H 0 t / T 2 n

e 0.4 10 V u F o [

2.0 c 0.2 200

2.5 0.0 400 101 4 6 8 10 12 2.5 2.0 1.5 1.0 0 5 10 15 20 RGC [Fe/H]

Figure 3. Left panel: The distribution of stars in RGC − [F e/H] plane with color-coded distance error. Middle panel: The rotational velocity distribution of the K giant stars is shown as the gray dots. The black solid lines in the blue and red shaded regions indicate the rotational velocities of the disk and halo components, respectively. The shadow colors from heavy to light represent the uncertainties from 1σ to 3σ. Right panel: The distribution of the errors of three velocity components. The red, green and blue arrows indicate the median errors for VR, VT and VZ , respectively. 4.1. Identification of the three components

(k) (k) +4 −1 The 1st component with VT,1 = 27−5 km s and pi(VT |fi,VT,i, σi,  ) = σ = 72+4 km s−1 is likely the stellar halo popula- f  (V k − V )2  (1) T,1 −4 i exp − T T,i , tion, since the dispersion is quite similar to Smith et p 2 2 2(σ2 + 2 ) 2π(σi + k) i k al.(2009) and Bird et al.(2018) at galactocentric dis- tance of 10 kpc. The 2nd component shows quite faster where i = 1, 2, 3 represent for the three components, fi, +6 −1 rotation with VT,2 = 182 km s and smaller disper- VT,i, and σi stand for the fraction, mean velocity, and −6 +3 −1 velocity dispersion for the ith component, respectively. sion σT,2 = 45−3 km s . It is noted that Morrison et al. (1990) claimed that the metal-weak thick disk has rota- k is the error of VT for kth star. According to Bayes’ −1 theorem, the posterior probability distribution of the tional velocity of < Vφ >= 170 ± 15 km s , Beers et al. unknown parameters can be written as (2014) also obtained quite similar values < Vφ >= 181 −1 −1 km s with a dispersion 53 km s and < Vφ >= 166 n 3 −1 −1 Y X (k) km s with a dispersion 47 km s using the metal p(θ1, θ2, θ3|VT ) ∝ pi(VT |θi,  )p(θ1, θ2, θ3), weak thick disk samples selected by Bidelman & Mac- k=1 i=1 (2) Connell(1973) and Ruchti et al.(2011) respectively, all of which are quite consistent with our results. Therefore, where θi = (fi,VT,i, σi) is the paramter vector for the ith component. k represents for the kth star in the the second component in our results is believed to be the samples with totally n stars. We adopt simplistic and metal-weak thick disk dynamically. The 3rd component +47 −1 loose uniformly distributed priors such that 0 < f < 1, shows retrograde rotation with VT,3 = −99−58 km s i +23 −1 −1 −1 and larger dispersion of 124 km s . As we see in −100 < VT,1 < 100 km s , 100 < VT,2 < 300 km s , −19 −1 −1 Section 4.3, this population has lower metallicity than −200 < VT,3 < 0 km s , and σi < 200 km s . We fur- the other two. It is either the so-called outer halo (Car- ther requires f1 + f2 + f3 = 1 so that f3 is not a free ollo et al. 2007) or an accreted debris in the solar neigh- parameter but derived from f1 and f2. Therefore, we borhood. In this work we focus on the stellar halo, i.e. totally have 8 free parameters in the model, i.e. f1, f2, the 1st component. VT,1, VT,2, VT,3, σT,1, σT,2, and σT,3. Then we apply emcee (Foreman-Mackey et al. 2013) to run a Markov Chain Monte Carlo (MCMC) simula- 4.2. The rotation of the local stellar halo tion with Affine solver for Eq. (2) with 2000 burn-in The stellar halo with the heliocentric distance of 4 iterations. Figure5 shows the results of the MCMC. kpc (the 1st component) shows a significant prograde +4 −1 We adopt the median values of the MCMC samples as rotation of 27−5 km s . Comparing to other works, the best-fit parameters and show the best-fit models in such rotation is different with the value from Deason the top-left panel of Figure4. The differences between et al.(2017) and the value for K giant samples from the median values and the 16% and 84% values are cal- Kafle et al.(2017). What should be kept in mind is that culated as the upper and lower uncertainties for each their samples have no overlapping with ours. Morrison parameter. The first row of Table1 lists the best-fit pa- et al.(1990) also studied the halo and disk with K giant rameters and their uncertainties of the three Gaussians. stars within 4 kpc and found that the stellar halo with [Fe/H]< −1.6 progradely rotates with 25±15 km s−1, 4. RESULT which is quite consistent with ours, while the dispersion Halo Rotation 5

200 ALL 60 -2.5<[Fe/H]<-1.6

150 40

100 Count Count 20 50

0 0

25 25 0 0 25 25 Data-Model Data-Model -400 -300 -200 -100 0 100 200 300 -400 -300 -200 -100 0 100 200 300 1 1 VT (km s ) VT (km s ) 80 60 -1.6<[Fe/H]<-1.3 -1.3<[Fe/H]<-1.0

60 40 40 Count Count 20 20

0 0

25 25 0 0 25 25 Data-Model Data-Model -400 -300 -200 -100 0 100 200 300 -400 -300 -200 -100 0 100 200 300 1 1 VT (km s ) VT (km s )

Figure 4. The top-left panel shows the distribution of VT for all the sample stars (black dots). The error bars are derived from Poisson distribution. The bin size is set at 10 km s−1. The red, blue, and green dashed lines indicate the best-fit Gaussian components i = 1, 2, and 3 respectively. The thick black solid line indicates the total distribution of the three-Gaussian model. The lower plot in the panel shows the residual distribution between the data and the model. No significant substructure is found in the residuals. The other panels are similar to the top-left one but for different [Fe/H] bins. From top-right to the bottom- right, the panels show the distribution of VT and the corresponding best-fit models with metallicity bins of -2.5<[Fe/H]<-1.6, -1.6<[Fe/H]<-1.3, and -1.3<[Fe/H]<-1.0, respectively. they detected is at 98±13 km s−1, slightly larger than is not significant. In light of those works, we detect this work. Considering the recent work by Belokurov et the rotational velocity for the sub-samples with different al.(2018), where the volumes selected overlap with ours, metallicities. the rotational velocity are quite similar with ours, 20 to We separate the samples into three sub-groups with 30 km s−1 in spherical coordinate. −2.5 <[Fe/H]< −1.6, −1.6 <[Fe/H]< −1.3, and −1.3 <[Fe/H]< −1.0. For each sub-sample, similar 4.3. VT versus metallicity MCMC simulation is applied and the best-fit 3-Gaussian Deason et al.(2017) used accurate proper motion to model parameters are listed in Table1. The comparison study the rotational velocity of halo stars and argued between the observed distributions of VT and the models that the prograde rotation of the metal richer K giants is for different metallicities are displayed in Figure4. faster than that of the metal poorer sample. Kafle et al. The black line with red shadow in Figure3 indicates (2017) also obtained a similar trend, but the difference VT,1 as a function of [Fe/H] for the halo population (the 6 Tian et al.

0.36 2

f 0.30

0.24

0.18

0.12 40 1

, 32 T

V 24

16

80 1

, 72 T

64

56

195 2 , T 180 V

165

60

2 54 , T 48

42

36 0

40 3 ,

T 80 V 120

160

180 3 ,

T 150

120 90

60 0 0.5 0.6 0.7 0.8 16 24 32 40 56 64 72 80 36 42 48 54 60 80 40 60 90 0.120.180.240.300.36 165 180 195 160120 120 150 180 f1 f2 VT, 1 T, 1 VT, 2 T, 2 VT, 3 T, 3

Figure 5. The results the MCMC simulation. The correlations of any two parameters are shown in the corner plot. The cyan lines are the median values of each individual parameters.

Table 1. Best fit parameters of 3-Gaussian Models derived from MCMC.

f1 f2 f3 VT,1 σT,1 VT,2 σT,2 VT,3 σT,3 N km s−1 km s−1 km s−1 km s−1 km s−1 km s−1 +0.038 +0.034 +4 +4 +6 +3 +47 +23 all 0.741−0.047 0.231−0.030 0.028 27−5 72−4 182−6 45−3 -99−58 124−19 3827 +0.089 +0.041 +9 +5 +12 +21 +32 +21 -2.5<[Fe/H]<-1.6 0.828−0.164 0.046−0.017 0.126 38−8 85−7 212−37 30−9 -47−58 113−9 1134 +0.066 +0.068 +7 +5 +18 +8 +60 +25 -1.6<[Fe/H]<-1.3 0.752−0.071 0.217−0.061 0.031 11−7 63−5 159−18 56−8 -115−55 158−26 1020 +0.058 +0.039 +7 +7 +6 +3 +23 +25 -1.3<[Fe/H]<-1.0 0.550−0.059 0.421−0.043 0.029 24−6 60−7 176−6 47−3 -31−52 112−9 1673 Halo Rotation 7

1st component). Rather than an increasing rotational a non-ignorable fraction of stars in the local volume, velocity with the metallicity argued by previous works, especially with metallicity around -1.5, what’s more, al- this result prefers to an essentially flat relation between most all of those stars are retrogradely rotating. This VT,2 and [Fe/H]. In other word, we do not find the rota- indicates that the relatively slower rotation velocity in tion of the local stellar halo is substantially correlated our results with metallicity between -1.6 and -1.3 might with metallicity. be caused, at least partly, by the major merger event. It is worthy to note that the K giant samples used by Deason et al.(2017) are within 50 kpc and |z| > 4 kpc 4.5. Impact from parameter choices and the samples used by Kafle et al.(2017) are within As the LSR velocity brings a directly change in the ro- 17 kpc and |z| > 4 kpc. Neither of them is spacially tational velocity during the conversion from proper mo- overlapping with our sample. tions, radial velocities and distances, we test the effect The fraction f1 of the halo population seems also in- of different choices of VLSR. We select three VLSR val- dependent of [Fe/H], as seen in the corresponding col- ues from literatures, 240 km s−1 (Reid et al. 2014), 232 −1 −1 umn of Table1. However, we find that f1 is peaked at km s (Sch¨onrich et al. 2010) and 220 km s (Dehnen -2<[Fe/H]<-1.6. We find that f , the fraction of the +4 +4 1 2000) for the test. We obtain that VT = 35−5, 27−5, halo population, decreases with metallicity, because of +4 −1 −1 and 15−5 km s for VLSR = 240, 232, and 220 km s , the increasing of the disk population. This again con- respectively. It is clear that a larger VLSR leads to a firms that the first component should be the stellar halo. higher halo rotation speed. However, even with the min- −1 Note that the selection effect in [Fe/H] is not taken into imum value of VLSR=220 km s , the stellar halo is still account. in substantially prograde rotation. Moreover, we also The black line in the blue shadow in Figure3 shows find that the rotational velocity is essentially indepen- rotational velocities at different metallicity bins for the dent of the metallicity with different VLSR. disk component (the 2nd component). It is seen that the uncertainties of VT,2 become smaller at metal richer 5. CONCLUSION side. This is because the fraction of the disk becomes We have studied the rotation of the halo with local larger when [Fe/H] increases, as seen in the column f2 K giant star samples within 4 kpc. Using the proper of Table1. Thus more stars are involved in this com- motions and radial velocities from Gaia and metallicities ponent. This supports that the component has higher from LAMOST, we draw the following conclusions. metallicity than the halo population. Firstly, the rotational velocity of the local stellar halo Finally, f column in Table1 shows that the 3 is strongly correlated to the LSR speed, VLSR, and also counter-rotational component occupies about 12.6% the azimuthal velocity of the Sun, V . With VLSR = 232 at −2.5 <[Fe/H]< −1.6 and only contributes 2.9% at −1 −1 km s and V =12.24 km s , the halo is progradely −1.3 <[Fe/H]< −1.0. This means that it is contributed +4 −1 rotating with VT = 27−5 km s . The dispersion of VT by the most metal-poor stars. The averaged metallic- +4 −1 of the halo K giant stars is σVT = 72−4 km s . ity of the component may be even lower than the halo Secondly, we find that the rotational velocity of the population. This is consistent with Carollo et al.(2007) stellar halo is independent of the metallicity in the local that the outer halo has metallicity of ∼-2.2 dex and in volume, which is different with the results in the outer counter-rotation. volume claimed by Deason et al.(2017) and Kafle et al.(2017). The flat relationship between VT and [Fe/H] 4.4. Possible mechanism of the prograde rotation of hints that the rotation may be due to secular rotation, the stellar halo rather than due to the net angular momenta from the minor mergers. The flat relationship between VT,1 and [Fe/H] implies that the prograde rotation may not be originated from Finally, we also identified a metal-poor and counter- the sustained angular momenta of the minor mergers, rotating hot component with rotational velocity of +47 −1 but in favor of an effect of secular evolution. Indeed, −99−58 km s , which is likely the outer halo. And +6 −1 Athanassoula et al.(2013) suggested that the rotating a disk-like component rotating with VT =182−6 km s bar in the central region of the Galactic disk may trans- is also identified, which is likely the metal-weak thick fer angular momenta to the halo. It is also noted that disk by comparing the results with previous results Xu et al.(2018) found the halo is oblate with axis-ratio (Morrison et al. 1990; Beers et al. 2014). lower than 0.5 at around 10 kpc, which may be related to the rotation of the halo. It is noted that, from Figure This work is supported by the National Key Ba- 2 in Helmi et al.(2018) the Gaia-Enceladus contributes sic Research Program of China 2014CB845700. CL 8 Tian et al. acknowledges the NSFC under grants 11873057 and nese Academy of Sciences. Funding for the project 11333003. The LAMOST FELLOWSHIP is supported has been provided by the National Development and by Special Funding for Advanced Users, budgeted Reform Commission. LAMOST is operated and man- and administrated by Center for Astronomical Mega- aged by the National Astronomical Observatories, Chi- Science, Chinese Academy of Sciences (CAMS). HT nese Academy of Sciences. This work has made use of is supported by the Young Researcher Grant of Na- data from the European Space Agency (ESA) mission tional Astronomical Observatories, Chinese Academy Gaia (https://www.cosmos.esa.int/gaia), processed by of Science. X.-X Xue thanks the ”Recruitment Pro- the Gaia Data Processing and Analysis Consortium gram of Global Youth Experts” of China and NSFC (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/ 11390371. Guoshoujing Telescope (the Large Sky Area consortium). Funding for the DPAC has been provided Multi-Object Fiber Spectroscopic Telescope LAMOST) by national institutions, in particular the institutions is a National Major Scientific Project built by the Chi- participating in the Gaia Multilateral Agreement.

REFERENCES

Athanassoula, E., Machado, R. E. G., & Rodionov, S. A. Gaia Collaboration, Helmi, A., van Leeuwen, F., et al. 2013, MNRAS, 429, 1949 2018, arXiv:1804.09381 Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Gaia Collaboration, Katz, D., Antoja, T., et al. 2018, Mantelet, G., & Andrae, R. 2018, arXiv:1804.10121 arXiv:1804.09380 Beers, T. C., Carollo, D., Ivezi´c, Z.,ˇ et al. 2012, ApJ, 746, 34 Gaia Collaboration, Spoto, F., Tanga, P., et al. 2018, Beers, T. C., Norris, J. E., Placco, V. M., et al. 2014, ApJ, arXiv:1804.09379 794, 58 Gaia Collaboration, Babusiaux, C., van Leeuwen, F., et al. Belokurov, V., Zucker, D. B., Evans, N. W., et al. 2006, 2018, arXiv:1804.09378 ApJL, 642, L137 Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. Belokurov, V., Erkal, D., Evans, N. W., Koposov, S. E., & 2018, arXiv:1804.09365 Deason, A. J. 2018, MNRAS, 478, 611 Grillmair, C. J. 2009, ApJ, 693, 1118 Grillmair, C. J., & Carlin, J. L. 2016, Tidal Streams in the Bernard, E. J., Ferguson, A. M. N., Schlafly, E. F., et al. Local Group and Beyond, 420, 87 2016, MNRAS, 463, 1759 Grillmair, C. J., & Dionatos, O. 2006, Bulletin of the Bidelman, W. P., & MacConnell, D. J. 1973, AJ, 78, 687 American Astronomical Society, 38, 48.03 Binney, J., & Tremaine, S. 2008, Galactic Dynamics: Grillmair, C. J., & Dionatos, O. 2006, ApJL, 643, L17 Second Edition, by James Binney and Scott Grillmair, C. J., & Dionatos, O. 2006, ApJL, 641, L37 Tremaine. ISBN 978-0-691-13026-2 (HB). Published by Hattori, K., Yoshii, Y., Beers, T. C., Carollo, D., & Lee, Princeton University Press, Princeton, NJ USA, 2008. Y. S. 2013, ApJL, 763, L17 Bird, S. A., Xue, X.-X., Liu, C., et al. 2018, Helmi, A., White, S. D. M., de Zeeuw, P. T., & Zhao, H. arXiv:1805.04503 1999, Nature, 402, 53 Carollo, D., Beers, T. C., Lee, Y. S., et al. 2007, Nature, Helmi, A., & de Zeeuw, P. T. 2000, MNRAS, 319, 657 450, 1020 Helmi, A., Babusiaux, C., Koppelman, H. H., et al. 2018, Bovy, J. 2015, ApJS, 216, 29 Nature, 563, 85 Cui, X.-Q., Zhao, Y.-H., Chu, Y.-Q., et al. 2012, Research Ibata, R. A., Gilmore, G., & Irwin, M. J. 1994, Nature, in Astronomy and Astrophysics, 12, 1197 370, 194 Deason, A. J., Belokurov, V., Koposov, S. E., et al. 2017, Johnson, D. R. H., & Soderblom, D. R. 1987, AJ, 93, 864 MNRAS, 470, 1259 Johnston, K. V., Zhao, H., Spergel, D. N., & Hernquist, L. Dehnen, W. 2000, AJ, 119, 800 1999, ApJL, 512, L109 Deng, L.-C., Newberg, H. J., Liu, C., et al. 2012, Research Kafle, P. R., Sharma, S., Lewis, G. F., & Bland-Hawthorn, in Astronomy and Astrophysics, 12, 735 J. 2013, MNRAS, 430, 2973 Foreman-Mackey, D., Hogg, D. W., Lang, D., & Goodman, Kafle, P. R., Sharma, S., Robotham, A. S. G., et al. 2017, J. 2013, PASP, 125, 306 MNRAS, 470, 2959 Frenk, C. S., & White, S. D. M. 1980, MNRAS, 193, 295 Liu, C., Xu, Y., Wan, J.-C., et al. 2017, Research in Gaia Collaboration, Eyer, L., Rimoldini, L., et al. 2018, Astronomy and Astrophysics, 17, 096 arXiv:1804.09382 Liu, C., Deng, L.-C., Carlin, J. L., et al. 2014, ApJ, 790, 110 Halo Rotation 9

Loebman, S. R., Valluri, M., Hattori, K., et al. 2018, ApJ, Sch¨onrich, R., Asplund, M., & Casagrande, L. 2011, 853, 196 MNRAS, 415, 3807 Mateu, C., Read, J. I., & Kawata, D. 2018, MNRAS, 474, Sch¨onrich, R., & Aumer, M. 2017, MNRAS, 472, 3979 4112 Sirko, E., Goodman, J., Knapp, G. R., et al. 2004, AJ, 127, McMillan, P. J. 2017, MNRAS, 465, 76 914 Morrison, H. L., Flynn, C., & Freeman, K. C. 1990, AJ, Smith, M. C., Evans, N. W., Belokurov, V., et al. 2009, 100, 1191 MNRAS, 399, 1223 Perryman, M. A. C., de Boer, K. S., Gilmore, G., et al. Sommer-Larsen, J., Beers, T. C., Flynn, C., Wilhelm, R., & 2001, A&A, 369, 339 Reid, M. J., Menten, K. M., Brunthaler, A., et al. 2014, Christensen, P. R. 1997, ApJ, 481, 775 ApJ, 783, 130 Tian, H.-J., Liu, C., Carlin, J. L., et al. 2015, ApJ, 809, 145 Rockosi, C. M., Odenkirchen, M., Grebel, E. K., et al. 2002, Tian, H.-J., Gupta, P., Sesar, B., et al. 2017, ApJS, 232, 4 AJ, 124, 349 Xu, Y., Liu, C., Xue, X.-X., et al. 2018, MNRAS, 473, 1244 Ruchti, G. R., Fulbright, J. P., Wyse, R. F. G., et al. 2011, York, D. G., Adelman, J., Anderson, J. E., Jr., et al. 2000, ApJ, 737, 9 AJ, 120, 1579 Sanderson, R. E., Helmi, A., & Hogg, D. W. 2015, ApJ, Zhao, G., Zhao, Y.-H., Chu, Y.-Q., Jing, Y.-P., & Deng, 801, 98 L.-C. 2012, Research in Astronomy and Astrophysics, 12, Sch¨onrich, R., Binney, J., & Dehnen, W. 2010, MNRAS, 723 403, 1829